Extensive literature documents increased incidence of depression in patients with diabetes mellitus. However, the degree to which the treatment of depression impacts on diabetes control is not clear. Some investigators have found a strong association between depression and diabetes control, while others have found a weak association. A review of existing literature as well as preliminary data from our laboratory suggest that this discrepancy may be due, in part, from failure of many studies to clearly differentiate type 1 and type 2 diabetes in their patient samples. The strongest association between depression and diabetes control has been reported in studies of patients with type 1 diabetes. Preliminary data from our group suggests that the Beck Depression Inventory (BDI) scores may be more significantly related to hemoglobin A1C (HbA1c) in type 1, than in type 2, diabetes mellitus. Studies assessing the effects of pharmacologic antidepressant therapy on diabetes have been confounded by the combined use of both diagnostic categories as well as by the direct metabolic effects of most antidepressant drugs. An emerging literature on the use of Cognitive Behavior Therapy (CBT) suggests that improving affect through behavioral intervention can improve diabetes control, particularly in type 1 patients. The overall aim of this proposal is to determine if CBT differentially improves glucose control in type 1 and type 2 diabetes patients. We hypothesize that CBT will produce a greater reduction in HbA1c in type 1 diabetes than in type 2 diabetes and that CBT-induced improvement in HbA1c is mediated by an improvement in depression. We will measure the effects of CBT on changes in HbA1c and daily blood glucose in a sample of 150 depressed patients with type 1 and type 2 diabetes over one year. To evaluate the mechanism by which CBT-induced changes in depression affect blood glucose, we will determine the role of cortisol and the role of diabetes self-care behaviors as mediating variables. Changes in cortisol and self-care are predicted to impact blood glucose levels to a greater extent in type 1 diabetes because these individuals are metabolically more sensitive to any variation.